Browse Articles

The intelligence illusion: why AI isn’t as smart as it is made out to be

Proteomic analysis identifies distinct signaling pathways in lumbar intervertebral disc degeneration between type-2 diabetic and non-diabetic patients

HIF1α mediates resistance to radiation and to KRAS inhibitors in pancreatic adenocarcinoma

Pancreatic ductal adenocarcinoma (PDAC) is highly treatment resistant and characterized by a hypoxic microenvironment. Here, we investigated the role of hypoxia-inducible factor 1α (HIF1α) in regulating resistance to radiation and KRAS-inhibitor. We employed CRISPR/Cas9 to knock out (KO) HIF1α from the murine KRAS G12D/+ ; p53 R172H/+ KPC and the KRAS G12D/+ ; p53 R273H ; CDK2NA -/- Panc-1 human pancreatic cell lines. Compared to WT, the HIF1α KO cell lines demonstrated a shift toward an epithelial phenotype and had decreased proliferation and migration under hypoxia. HIF1α KO cell lines were less likely to survive after radiotherapy, and neutral comet assays demonstrated DNA damage four hours after treatment, suggesting that HIF1α promotes radioresistance through non-homologous end joining. When treated with a KRAS G12D inhibitor, HIF1α KO cells exhibited significantly increased apoptosis due to decreased p53 degradation, likely mediated through Mdm2. Confirming this, enrichment of hypoxic signaling was associated with KRAS inhibitor resistance in a cohort of 31 KRAS G12D cell lines. Our results thus suggest that inhibiting HIF1α may sensitize PDAC to radiation and KRAS inhibitors. To explore this, we conducted a drug repurposing screen and identified three HIF1α inhibitors (bakuchiol, BAY-87–2243, 2-methoxyestradiol) whose sensitivities were correlated with sensitivity to Deltarasin, a KRAS inhibitor. Our findings suggest that HIF1α inhibitors could be used to sensitize PDAC to radiotherapy and KRAS inhibitors.

Forty-five years of progress after a key paper about the evolution of cooperation

A fatigue driving detection method based on driver posture and facial state analysis

Plant trait diversity buffers soil moisture dynamics on coastal dikes during drought periods

Soil moisture is considered a key component for the structural integrity of engineered ecosystems, such as sea dikes. Although plants are important determinants of physical soil properties in dike greening, research lacks on the extent to which greater biodiversity can mitigate soil moisture loss during extreme weather events. This provided the motivation to investigate the influence of two plant communities of different species composition – namely, an herb-dominated vegetation area (‘Mix-Herb’) compared to a grass-dominated area (‘Mix-Grass’) – on soil physical conditions over the course of one year on a summer dike in northern Germany. Vegetation mapping, high-resolution measurements of soil temperature and moisture, and comprehensive precipitation data provided the framework for the investigations. It was found that species diversity (Shannon Index) declined over time from 2.7 to 2.3 for ‘Mix-Herb’ and from 2.2 to 2.0 for ‘Mix-Grass’. In-situ measurements of soil physical conditions revealed that the ‘Mix-Herb’ plant community moderated diurnal soil temperature variations more effectively than ‘Mix-Grass’. During a drought in June 2023, the ‘Mix-Herb’ vegetation area was also considerably less affected by soil heating and moisture deficit. However, after mowing, the thermal buffer effect reversed and greater diurnal temperature variations occurred in the soils of the herbaceous vegetation. During a second drought in September 2023, the’Mix-Grass‘soils exhibited higher moisture loss rates after mowing. These findings highlight the importance of the functional composition of plant communities and management practices such as mowing schedules, tailored spatially and temporally to ecological and climatic conditions, for regulating the soil microclimate on dike systems, with potential implications for dike’s resistance under climatic extremes.

How the idea of human superiority over nature was invented

Removal of antibiotics and antibiotic resistance genes from domestic wastewater using mesocosm-scale constructed wetlands with different filter media

Enhancing autonomous agriculture control systems in greenhouses for sustainable resource usage using deep learning techniques

Greenhouse climate control is essential for optimizing crop growth while minimizing resource consumption in controlled environment agriculture. Traditional rule-based and fixed-action strategies often struggle to achieve a balance between these objectives. This paper proposes a reinforcement learning (RL) based framework for greenhouse climate control, integrating deep learning models to predict both crop growth and resource consumption. The framework enables an RL agent to optimize greenhouse control setpoints dynamically, maximizing crop yield while ensuring sustainable resource usage. The proposed system incorporates a Multi-Layer Perceptron (MLP) model to predict internal greenhouse climate conditions, a Long Short-Term Memory (LSTM) model for crop parameter estimation, and a separate LSTM model for forecasting daily resource consumption. These models collectively simulate a greenhouse environment where an RL agent learns to regulate temperature, CO 2 concentration, and irrigation levels by interacting with the virtual environment. A custom reward function is designed to guide the agent, considering key crop parameters; stem elongation, stem thickness, and cumulative trusses; alongside resource consumption metrics, including heating, electricity, CO 2 , and irrigation costs. To enhance the adaptability of the RL agent, a feature-selection mechanism identified the most influential climate and control features, reducing observation complexity and accelerating convergence. Retraining under stochastic weather conditions strengthened robustness to dynamic environments, enabling the agent to consistently outperform fixed-action strategies. Evaluation revealed a stable Pareto frontier between yield and resource consumption, confirming that the framework accurately captured the productivity and sustainability trade-off and remained robust across varying reward-weight settings. Comparative analysis of multiple RL algorithms; Proximal Policy Optimization (PPO), Deep Deterministic Policy Gradient (DDPG), Soft Actor-Critic (SAC), and Twin Delayed Deep Deterministic Policy Gradient (TD3) demonstrated that TD3 outperforms other algorithms, achieving the highest cumulative rewards and reaching optimal policies faster. Experimental evaluations demonstrate that the proposed TD3 RL-based greenhouse control system achieves higher crop yield growth rates while optimizing resource usage, outperforming conventional greenhouse control strategies. This study presents a novel data-driven, adaptive greenhouse management approach, bridging the gap between crop growth modeling and autonomous climate control, contributing to sustainable and intelligent agricultural practices.

How pain intensity and mental disorders shape chronic pain sick leave and quality of life in the general Spanish population

Genetic variations associated with immediate hypersensitivity reactions to iodinated contrast media: A whole exome sequencing study

Objective The use of iodinated contrast media (ICM) in computed tomography (CT) has increased significantly; however, hypersensitivity reactions (HSRs) remain a concern. This study aimed to investigate genetic factors associated with ICM-induced immediate HSRs using whole exome sequencing (WES). Materials and Methods We conducted a case–control study including 20 patients with ICM-induced immediate HSRs and 11 controls who had received ICM at least three times without HSRs. WES was performed with DNA extracted from saliva samples. Analyses included single-nucleotide variant (SNV) association testing using the Cochran–Armitage trend test with false discovery rate (FDR) correction, gene-wise variant burden (GVB) analysis, and copy number variation (CNV) detection using complementary algorithms. Results A variant in FAST kinase domain 1 ( FASTKD1 , rs12618227) was significantly more prevalent in the control group compared with the case group (72.7% vs. 5.0%, FDR p  < 0.10), suggesting a protective role. GVB analysis revealed lower scores for FASTKD1 and 2-hydroxyacyl-CoA lyase 1 ( HACL1 ) in the control group (nominal p  < 0.001). CNV analysis identified a significant Signal Regulatory Protein Beta 1 ( SIRPB1 ) deletion in the case group (5/20, 25.0%). In contrast, CNVs in Mucin 12, cell surface associated ( MUC12 ) were observed in both groups. Immune cell expression data showed high expression of FASTKD1 , HACL1 , and SIRPB1 in granulocytes, particularly basophils. Conclusion FASTKD1 and HACL1 , which are involved in mitochondrial and metabolic regulation, and SIRPB1 , which participates in innate immune signaling, were identified as candidate genes potentially associated with ICM-induced immediate HSRs. These suggest a possible contribution of both metabolic and immune regulatory pathways to genetic susceptibility and require validation in larger, independent cohorts before clinical application.

Drowning in data sets? Here’s how to cut them down to size

Exploring the limits of pre-trained embeddings in machine-guided protein design: a case study on predicting AAV vector viability

Abstract Effective representations of protein sequences are widely recognized as a cornerstone of machine learning-based protein design. Yet, protein bioengineering poses unique challenges for sequence representation, as experimental datasets typically feature few mutations, which are either sparsely distributed across the entire sequence or densely concentrated within localized regions. This limits the ability of sequence-level representations to extract functionally meaningful signals. In addition, comprehensive comparative studies remain scarce, despite their crucial role in clarifying which representations best encode relevant information and ultimately support superior predictive performance. In this study, we systematically evaluate multiple ProtBERT and ESM2 embedding variants as sequence representations, using the adeno-associated virus capsid as a case study and prototypical example of bioengineering, where functional optimization is targeted through highly localized sequence variation within an otherwise large protein. Our results reveal that, prior to fine-tuning, amino acid–level embeddings outperform sequence-level representations in supervised predictive tasks, whereas global sequence-level embeddings tend to be more effective in unsupervised settings. However, optimal performance is only achieved when embeddings are fine-tuned with task-specific labels, with sequence-level representations providing the best performances. Moreover, our findings indicate that the extent of sequence variation required to produce notable shifts in sequence representations exceeds what is typically explored in bioengineering studies, showing the need for fine-tuning in datasets characterized by sparse or highly localized mutations.

The relationship between live streamers’ self-disclosure and consumers’ purchase intention: The parallel mediating role of psychological distance and perceived homophily

As live streaming commerce continues to expand rapidly, the role of live streamers has emerged as a crucial factor. In an online environment with increasingly fierce competition and increasingly rich emotional needs of consumers, live streamers have gradually shown more self-disclosure to attract more consumers’ attention. While existing studies have indicated that the communication styles adopted by live streamers constitute a key factor influencing consumers’ purchasing behaviors, little is known about live streamers’ self-disclosure, a specific communication mechanism, is associated with consumers’ purchase intention, as well as the roles that psychological distance and perceived homophily between consumers and live streamers play in this process. Drawing on Social Penetration Theory, this study investigates the association between live streamers’ self-disclosure and consumers’ purchase intention, and analyzes the parallel mediating roles of psychological distance and perceived homophily in this relationship. The investigation employs regression analysis to examine 306 survey responses gathered from Chinese consumers. The results reveal that live streamers’ self-disclosure is positively associated with consumers’ purchase intention by diminishing the psychological distance between live streamers and consumers and enhancing perceived homophily; notably, psychological distance and perceived homophily exert significant parallel mediating effects in the association between live streamers’ self-disclosure and consumers’ purchase intention. Live streamers’ self-disclosure is linked to reduced audience psychological defensiveness via increased closeness, and to stronger identification through heightened similarity perception; both pathways are equally important in relation to higher purchase intention. This study contributes to a better understanding of how live streamers’ self-disclosure in live streaming contexts relates to consumers’ purchase intention, as well as the roles of psychological distance and perceived homophily between consumers and live streamers. It also provides a theoretical basis and practical implications for live streamers to engage in positive self-disclosure, to build closer connections and identification with consumers by reducing psychological distance and enhancing perceived homophily, and ultimately to support higher consumers’ purchase intention as well as the competitiveness of e-commerce platforms and enterprises.

Integrated wet lab and in silico discovery and characterization of bacteriophages with antibiotic synergy against multidrug resistant Acinetobacter baumannii

High-throughput high content quantification of HIV-1 viral infectious output

Infection with human immunodeficiency virus (HIV-1) remains a global health issue and still drives the development of significant pathology and various comorbidities. Antiretroviral therapy (ART) can effectively suppress viral replication but is often initiated months or years after initial infection, leaving a substantial period in which viral replication progresses unchecked. While ART suppresses HIV-1 replication, it does not prohibit the development of HIV-1-associated comorbidities, highlighting a lack of understanding in the connection between replication and HIV-1-associated pathogeneses. Further, a high percentage of all HIV-1 virions produced are non-infectious, and this proportion is much higher in ART-treated individuals, showing that despite inefficient viral replication, which becomes even less efficient with ART, HIV-1 is still able to drive disease. Thus, it is critical to better define HIV-1 replication dynamics to more effectively target different stages of the viral replication cycle in distinct cell populations. Here, we show a high-content imaging reporter assay that uses modified human osteosarcoma cells expressing HIV-1 receptors (GHOST cells) which fluoresce in response to HIV-1 infection. These cells have been previously used to assess HIV-1 infectivity and tropism, but this modified assay enables rapid evaluation of large numbers of samples with consistency and replicability, while also easily integrating into existing experimental pipelines that analyze p24 secretion in collected supernatants. This also allows for direct correlation between infectivity and p24 secretion, resulting in a deeper interrogation and more robust understanding of HIV-1 infection kinetics.

Integrated DFT, molecular docking, and molecular dynamics investigation of some novel 2-thiohydantoin analogues as potent CDK2 inhibitors for anticancer therapy

Abstract Cancer progression is driven by dysregulation of cyclin-dependent kinase 2 (CDK2), a critical cell cycle regulator. This study employed an integrated computational approach combining Density Functional Theory (DFT), molecular docking, molecular dynamics (MD) simulations, and MM-PBSA calculations to evaluate 2-thiohydantoin derivatives as CDK2 inhibitors. DFT calculations revealed compounds 2b-e narrowest lowest unoccupied molecular orbital (LUMO)- highest occupied molecular orbital (HOMO) gaps (3.02–3.26 eV in DMSO) and highest electrophilicity indices (> 3.20 eV), indicating enhanced reactivity toward biological targets. QTAIM and Fukui function analyses identified key electrophilic centers (C2, O12, C14) and hydrogen bonding sites essential for protein interactions. Molecular docking against CDK2 (PDB: 1HCK) showed compounds 2c , 2d , and 2b exhibited superior binding affinities (-9.312, -9.303, and − 9.269 kcal/mol) compared to ATP (-8.460 kcal/mol), forming critical hydrogen bonds with Lys33 and Thr14. The 10 ns MD simulations confirmed stable binding, with compound 2f maintaining highest conformational stability (RMSD ~ 0.05 nm) and robust hydrogen bonding (mean: 2.70 bonds). MM-PBSA analysis revealed compound 2d achieved optimal binding affinity (ΔG_bind = -34.50 ± 0.42 kcal/mol) through balanced van der Waals interactions (-50.74 kcal/mol) and minimal desolvation penalty (52.40 kcal/mol). Compounds 2b , 2c , 2d , and 2f emerged as lead candidates for experimental validation as next-generation CDK2-targeted anticancer agents.

Digital financial inclusion and rural income convergence in China: Evidence from household panel data

The rapid expansion of digital financial inclusion (DFI) in China has raised concerns that a rural-urban “digital divide” could worsen income inequality. Using 55,684 household-year observations from the China Family Panel Studies (2012−2022) combined with a provincial DFI index, we investigate whether DFI expansion helps rural households catch up or leaves them further behind. DFI is a significant driver of rural income growth. A one-standard-deviation increase in the DFI index raises per capita household income by 4.6%. Importantly, we find evidence of income convergence rather than divergence, as DFI benefits extend to both the bottom-40% and top-20% of income earners. However, these benefits are not evenly distributed. The positive effects of DFI are significantly stronger for households in more developed areas, those with higher education levels, and those headed by males. Mediation analysis reveals that DFI’s income effects operate through multiple complementary channels: while improved financial inclusion (asset ownership and credit access) plays a significant role, the majority of benefits flow through alternative pathways including e-commerce participation, reduced transaction costs for remittances, access to market information, and digital network effects. While DFI helps narrow the primary urban-rural gap, our findings highlight a new “digital-within-digital” divide. Policies promoting digital literacy and infrastructure in inland regions are crucial to prevent some rural subgroups from being left behind and to ensure equitable growth.

Novel approaches for understanding and improving the effectiveness of seed biopriming

Predictors of saccadic reaction time among young children in Lusaka, Zambia

Saccadic reaction time (SRT), an assessment of visual processing speed, may afford an accurate and unbiased measure of early childhood development (ECD). Few studies have examined SRT in low- and middle-income countries (LMICs), including its drivers. We sought to identify predictors of SRT as well as to assess the correlation between SRT and concurrent measures of ECD [Global Scales of Early Development (GSED) development-for-age Z-score (DAZ), height-for-age Z-score (HAZ), and head circumference-for-age Z-score (HCZ)], among young children in Lusaka, Zambia. We conducted a sub-study among 299 Lusakan children participating in a 2x2 cluster-randomized trial. SRT was assessed at ~31 months using a screen-based setup with a Tobii Pro Fusion tracker. Associations with household, caregiver, and child characteristics were assessed using univariable regression models; predictors significant at the p < 0.20 level were retained in a multivariable model. Pearson correlation coefficients were calculated to assess associations between SRT and other concurrent measures of ECD. In the multivariable model, characteristics found to be significant predictors of SRT included: being the only child <5 in the household at baseline (β: −10.02, 95% CI: −19.71, −0.33, p = 0.04), length-for-age Z-score (LAZ) at baseline (β: −3.17, 95% CI: −6.31, −0.04, p = 0.047), consuming ≥4 food groups in the past day (β: −10.42, 95% CI: −19.98, −0.86, p = 0.03), and having diarrhea in the past 2 weeks (β: 12.38, 95% CI: 0.71, 24.06, p = 0.04). SRT was significantly negatively correlated with HAZ (−0.176, p < 0.01) and HCZ (−0.132, p < 0.05), but not GSED DAZ. Overall, we identified several significant predictors of SRT among young children in Lusaka, Zambia, including birth spacing, baseline LAZ, dietary diversity, and diarrheal disease. Further research is needed, including in different age groups and geographic locations, to better understand the drivers of slow SRT, and poor ECD generally, in LMICs.